Molecular characteristics of therapy-related myeloid neoplasms in PARP inhibitor–treated patients.

6597 Background: Poly ADP-ribose polymerase inhibitors (PARPi) are part of standard therapy in solid tumors, and is associated with development of therapy-related myeloid neoplasm (t-MN), including myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). The contribution of pre-existing clonal hematopoiesis (CH) in t-MN with PARPi exposure is unknown. We report outcomes of a real-world cohort of t-MN patients following PARPi therapy. Methods: We performed a retrospective single-center analysis of 23 patients with solid tumors treated with PARPi (olaparib, rucaparib, niraparib, talazoparib) who subsequently developed t-MN. Chart review was performed to collect clinical data. Blood samples collected as a somatic control prior to the first chemotherapy or radiation for targeted next-generation sequencing (NGS) using the MSK-IMPACT platform were analyzed for pre-existing CH in a subset of evaluable patients (n=17). CH was defined as a somatic mutation with variant allelic frequency of 2% or more in peripheral blood. Results: Median age was 63 years at initial solid tumor diagnosis and 68 years at t-MN diagnosis. Primary malignancies included Ovarian/Fallopian Tube (83%, n=19), Breast (17%, n=4), Prostate (9%, n=2), and Esophageal (n=1) cancer. 3 patients had a history of both Breast and Ovarian cancer. Germline BRCA1/2 mutations were present in 48% (n=11) and 1 patient had Li-Fraumeni. Median time from PARPi initiation to t-MN diagnosis was 23.7 months (range: 1.6–45.3 months). At t-MN diagnosis, 52% (12/23) presented with MDS (including one case of clonal cytopenias of undetermined significance) and 48% (11/23) with AML. Genomic profiling of t-MN showed 78% (18/23) of patients had TP53 mutations, and 65% (15/23) had complex karyotype. 90% of evaluable AML cases were classified as adverse risk by ELN 2022 and most MDS cases (9/11) had High/Very High risk by IPSS-R. Median overall survival in this cohort was 6 months from t-MN diagnosis. Among 17 patients with evaluable baseline blood samples, 41% (7/17) had detectable CH prior to t-MN, while 59% (10/17) did not. Of those with pre-existing CH, 57% (4/7) harbored mutations in DNA damage response genes (2 with TP53 , 2 with PPM1D ). Other mutations included DNMT3A (N=2), ASXL1 (N=1), SRSF2 (N=1), and JAK2 (N=1). 43% (3/7) had a single CH mutation, while 57% (4/7) had 2 or more CH mutations. 6/7 of those with pre-existing CH had molecular testing at t-MN diagnosis, all of which (6/6) had a detectable TP53 mutation at t-MN diagnosis, even when not previously detected. Cytogenetic features of t-MN for those with and without pre-existing detectable CH were similar. Conclusions: This is the largest single center cohort of post-PARPi t-MN with molecular characteristics described to date. Further study on the effect of PARPi therapy in the pathogenesis of t-MN is needed.